mark j. moretto, lori m. feaga, michael f. a’hearn, silvia ... · pdf filenaturally...
Post on 07-Feb-2018
218 Views
Preview:
TRANSCRIPT
Naturally Occurring Outbursts at Comet Tempel 1 Mark J. Moretto, Lori M. Feaga, Michael F. AHearn, Silvia Protopapa, Jessica M. Sunshine, Tony L. Farnham
University of Maryland, College Park
The Deep Impact Mission Deep Impact encountered comet
Tempel 1 on July 4th 2005
The flyby spacecraft carried 2 visible imagers (HRI-Vis & MRI) and 1 infrared spectrometer (HRI-IR)
The impactor spacecraft carried 1 visible imager (ITS)
The impactor had a mass of 372 kg and collided with Tempel 1 at 10.3 km/s (AHearn et al., 2005)
This collision delivered 19 GJ of kinetic energy, the equivalent of 4.5 tons of TNT (AHearn et al., 2005)
Data was collected both pre and post-impact Fig. 1: The Deep Impact spacecraft
Research Goals Use HRI-IR spectra to determine if/how the composition of the coma changes due to an outburst
Determine the spatial distribution of volatiles before and after the outburst Make inferences about the cause of outbursts and comets in general
Deep Impacts Infrared Spectrometer HRI-IR operates between wavelengths of 1.05 and 4.85 microns, a region
where H2O, CO2, CO, and organic molecules have emission lines
Has a minimum spectral resolving power (/) of 200
Is a long-slit spectrometer, so each frame has a spatial dimension (512 pixels, unbinned) and a wavelength dimension(1024 pixels, unbinned)
Binned pixels have a FOV of 10-10 steradians The middle third of the slit is covered by an anti-saturation filter (ASF) to
prevent the nucleus from saturating the detector
Scans were used to gain a second spatial dimension Note that for these data the pixels are twice as large in the scan direction
as they are in the slit direction due to the imaging mode, BINFF
Data
Each scan consists of 50 8-second frames and was acquired in unbinned full frame mode. The comet is located inside the ASF for these data.
Scan ID UT Time on 2 Jul 2005 (H:M:S)
Time relative to Outburst
Abbreviation
8600000 06:00:22.153 -2.5 hours 7_2_0
8600001 07:54:32.105 -0.5 hours 7_2_1 (Pre-Outburst)
8600002 10:00:22.057 +1.5 hours 7_2_2 (Post-Outburst)
8600003 12:05:12.110 +3.5 hours 7_2_3
Data Calibration Data are decompressed, if needed
Linearity coefficients are applied A dark frame is scaled to fit the scan, then is subtracted from the data A flat-field is divided out to correct for pixel to pixel variations in sensitivity Our flat-field also corrects for the transmission profile of the ASF Data are converted to radiance units and a wavelength value was
assigned to each pixel
Creating a Viable Dark A dark made from the last 5 frames of the scan being calibrated was
determined to be inadequate
These darks did not allow for a proper analysis of the long wavelength end of the spectrum and the dark created for scan 8600002 introduced artifacts into the data
A dark was created by averaging the last 5 frames from scans 7_2_0, 7_2_1, and 7_2_3
Outliers beyond 2.5 were excluded from the mean An optimal scaling factor was determined by minimizing the average 2 of
the best continua that could be fit to the data, though discretion was used
Outbursts at Tempel 1 12 were observed by Deep Impact (AHearn et
al., 2005, McLaughlin, private communication)
1 was observed by the Calar Alto observatory and Hubble Space Telescope (Lara et al., 2006, Feldman et al., 2007)
Large outbursts eject ~106 kg of material (Belton et al., 2008)
Outbursts occur at a rate of ~0.3 per day, with some rotational phases more likely to produce an outburst than others (Farnham et al., 2007)
Outbursts are directional and vary in strength (Farnham et al., 2007)
Only the 2 Jul 2005 outburst was large enough to be observed by HRI-IR
These outbursts occur too frequently to be the result of impacts
Fig. 3: Pre-impact photometry from Deep Impact that shows 2 outbursts occurring at similar rotational phases
(AHearn et al., 2005)
Data Reduction A spectrum is created by totaling spectra from different spatial coordinates
The spectrum is fitted with a continuum, which is modeled using a smoothed solar spectrum and a Plank function (numerous solutions are tested and a chi-squared test determines the best fit)
Once the continuum is subtracted, emission bands can be integrated over and converted to the number of molecules in the field of view
Results for Full Apertures Nucleus-centered square apertures (in pixel space) were created and a single continuum was fit to
the totaled spectrum then the abundance of water vapor and CO2 were plotted versus aperture size for the scans bracketing the outburst
It is clear that the abundance of water vapor does not change as a result of the outburst There is an increase in the abundance of CO2 after the outburst and the magnitude of this increase
is dependent on aperture size, which is consistent with CO2 being released during an outburst
Duplication of Results: Shells Shells were created by taking the total spectrum of a square ring of spectra
A continuum was fit to each shell to get a well defined shape, then divided by the number of spatial pixels in the shell and normalized to each spectrum. Water vapor and CO2 surface brightnesses were then calculated.
A standard deviation was calculated for the water vapor and CO2 surface brightnesses for each shell. This was taken as the error for each pixel in that shell
These shells were used to recreate the total apertures by summing the surface brightnesses for water vapor and CO2 from the normalized spectra
Duplication of Results: Spectral Maps Spectral Maps were created by summing the spectra in a square box to get the shape of the
continuum, then normalizing the average continuum to each individual spectrum in the box
The sum of the surface brightnesses from the normalized spectra were used as the value for the central pixel of that box
Standard deviations were calculated for each continuum that was fit and for each pixel, as each pixel would have between 1 and the box size squared different continua fit to it
To recreate the full apertures the spectral maps were divided by the box size squared and then the surface brightnesses within each aperture were summed
Errors were calculated using both the standard deviation from the continuum and the standard deviation of each pixel, though the latter is about an order of magnitude smaller than the former
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0 5 10 15 20 25 30 35
Carbon
Dioxide
Surface Brig
htne
ss (W
/m^2/sr)
Aperture Size (Pixels)
Carbon Dioxide Surface Brightness
CO2 SHELL POST
CO2 SHELL PRE
CO2 POST
CO2 PRE
CO2 3*3 BOX MAP POST
CO2 3*3 BOX MAP NO NEGATIVES POST
CO2 3*3 BOX MAP PRE
CO2 3*3 BOX MAP NO NEGATIVES PRE
CO2 7*7 BOX MAP POST
CO2 7*7 BOX MAP PRE
Post-Outburst Cluster
Pre-Outburst Cluster
30
35
40
45
50
55
60
65
1.0008 1.0009 1.001 1.0011 1.0012 1.0013 1.0014 1.0015
Average Ch
i Squ
ared
Scaling Factor
Scaling Factor DeterminaEon: 8600001
Fig. 2: Average chi squared values for the best fit continua for aperture sizes of 1*1 pixels to 29*29 pixels versus the dark scaling factor used to calibrate the data.
Fig. 5: Carbon dioxide surface brightness versus aperture size pre and post-outburst with 1 error bars that were calculated for each data set.
Error Analysis Error bars calculated pre and post-outburst are consistent with each other
Changes in surface brightness are the main cause of changes in the SNR The CO2 surface brightness is larger post-outburst when compared to pre-
outburst for all aperture sizes, though this is more significant at larger aperture sizes
The CO2 surface brightness before the outburst becomes overwhelmed by noise at the 15*15 pixel aperture size, as can be seen by the surface brightness leveling off in Figure 5.
The CO2 surface brightness after the outburst does not appear to level off within a 35*35 pixel box, though an increase in the scatter between the different methods indicates that noise is becoming more significant at larger apertures, beyond roughly 19*19 pixels.
0
2
4
6
8
10
12
14
16
18
20
0 5 10 15 20 25 30
Signal to
Noise RaE
o
Aperture Size (Pixels)
Signal to Noise RaEos
SNR H2O Post
SNR H2O Pre
SNR CO2 Post
SNR CO2 Pre
Fig. 6: Signal to noise ratio versus aperture size for water vapor and CO2.
Percent Change The percent change in water vapor is negligible and within the error bars
CO2 is consistently more abundant after the outburst The increase in CO2 post-outburst is consistently ~50 % within the 15*15 pixel
aperture
If negative surface brightnesses are neglected, the percent increase in CO2 remains relatively constant
0
50
100
150
200
250
300
350
400
0 5 10 15 20 25 30 35
Chan
ge in Carbo
n Dioxide Surface Brightne
ss (%
)
Aperture Size (Pixels)
Percent Change in Carbon Dioxide Surface Brightness
% Change in CO2 3*3 Map, No NegaPves
% Change in CO2 3*3 Map
% Change in CO2 Shells
Fig. 7: Percent change
top related